Gene Ontology annotation: Worked
examples
Prudence Mutowo
UniProt-GOA
Talk Overview
• Intro to GO and GO terms
• Exercise
• Annotating to GO
• Accessing GO annotations
• Exercise
• Practical use of GO
• Exercise
• Precautions
QuickGO
QuickGO
Search field
Global tool bar
Term basket
QuickGO
QuickGO
QuickGO
Search whole proteomes
•
•
•
•
Gene annotation
What information can GO capture?
What information can GO not capture?
How specific can the annotations be?
How many annotations can be derived from a single
paper?
• How many annotations can be made to a single gene
product?
• Repetitive annotations?
• Conflicting annotations?
Worked example 1
• Foxc1, Foxc2
• Look at the figure legend
Foxc1
Foxc2
What is the difference between A and C?
Whole-mount in situ
hybridization of 9.5 dpc
embryos
anti-digoxigenin
antibodies
Transcripts in presomitic
mesoderm and somites
Whole-mount
immunohistochemistry
of 9.5 dpc
anti-Foxc1/c2 antibodies
Proteins in presomitic
mesoderm
What can we annotate from this?
Whole-mount in situ
hybridization of 9.5 dpc
Proteins may not stay in the same cell
embryos
Transcripts in presomitic as the transcript
mesoderm and somites
Whole-mount
immunohistochemistry
of 9.5 dpc
No tissue types in GO
Proteins in presomitic
Cellular component is in GO
mesoderm
Antibody stained dorsal
aorta (12.5 dpc embryo)
Proteins localised in
endothelial and smooth
muscle cells
Immunostaining of
human aortic smooth
muscle cells
Nuclear localisation of
proteins
✓
No cell types in GO
Cellular component is in GO
Cellular component is in GO
Nucleus (singular not plural)
Protein
GO term
Foxc1
GO:0005634 nucleus (cellular component)
Foxc2
GO:0005634 nucleus (cellular component)
What can we annotate from this?
Immunostaining of
human aortic smooth
muscle cells
Nuclear localisation of
proteins
Protein GO term
GO:0005634
Foxc1
nucleus
GO:0005634
Foxc2
nucleus
Supporting evidence
Figure 1G immunostaining shows
nuclear location
Figure 1H immunostaining shows
nuclear location
Annotation
example
2
Figure legend:
FIG. 2.
Human Nbp35 is a cytosolic protein.
Annotation
example
2
Human Nbp35 is a cytosolic protein.
Protein GO term
Supporting evidence
NUBP1
Fig 2A Immunofluorescence and/or
Fig 2C subcellular fractionation
GO:0005829
Cytosol
Annotation example 3
Bcl-XL/S: marker for
outer mitochondrial
membrane.
HAX1 annotation:
GO term
Supporting evidence
GO:0005739 mitochondrion
Fig 7a, b and d, co-localization of HAX1 and BclXL/S by immunofluorescence confirms localisation of
HAX1 to mitochondria
Or GO:0005741 mitochondrial
outer membrane
GO
has are
22 different
evidence
codes
What
Evidence
codes?
IEA, Inferred from Electronic Annotation
IDA, Inferred from Direct Assay
IMP, Inferred from Mutant Phenotype
IPI, Inferred from Physical Interaction
TAS, Traceable author statement
NAS, Non-traceable author statement
ISS, Inferred from Sequence Similarity
Experimental
evidence,
Methods &
Results
Abstract &
Introduction
Why use Evidence codes?
Users can quickly see how reliable the annotation is likely
to be
Eg. An author statement may suggest that the function of
the protein has not been demonstrated in your species.
Users can filter on specific evidence types
Eg. only using manual annotations in their analysis
Worked example 1
Immunostaining of
human aortic smooth
muscle cells
Nuclear localisation of
proteins
Protein
Foxc1
Foxc2
GO term
GO:0005634
nucleus
GO:0005634
nucleus
IDA: inferred from Direct Assay
IMP: inferred from Mutant Phenotype
Which evidence code?
IDA
IDA
Evidence codes (page 9)
IDA Inferred from direct assay
IMP Inferred from mutant phenotype
Is the experiment demonstrating the normal function of the
gene product?
eg immunofluorescence, transfection, enzyme assay
IDA
Is the abnormal expression of a gene or a mutation in a
protein indicating something about the ‘normal’ function?
eg knockout mouse, siRNA, overexpression
IMP
Example 2, which evidence code?
Choose evidence codes for examples 2 and 3 – 10 minutes
Human Nbp35 is a cytosolic protein.
Protein GO term
Supporting evidence
NUBP1
Fig 2A Immunofluorescence and/or
Fig 2C subcellular fractionation
GO:0005829
Cytosol
Inferred from Direct Assay (IDA)
• Immunofluorescence and subcellular fractionation cellular
component annotations
• IDA – inferred from direct assay
Use IDA
eg
Protein
GO term
1
Foxc1
GO:0005634 nucleus
IDA
1
Foxc2
GO:0005634 nucleus
IDA
2
Nbp35
GO:0005829 cytosol
IDA
Evidence code
Example 3, which evidence code?
Bcl-XL/S: marker for
outer mitochondrial
membrane.
HAX1 annotation:
GO term
Supporting evidence
GO:0005739 mitochondrion
Fig 7a, b and d, co-localization of HAX1 and BclXL/S by immunofluorescence confirms localisation of
HAX1 to mitochondria
Or GO:0005741 mitochondrial
outer membrane
Inferred from Direct Assay (IDA)
• Immunofluorescence and subcellular fractionation cellular
component annotations
• IDA – inferred from direct assay
• EXP – inferred from EXPeriment
eg
Protein
GO term
1
Foxc1
GO:0005634 nucleus
IDA
1
Foxc2
GO:0005634 nucleus
IDA
2
Nbp35
GO:0005829 cytosol
IDA
3
HAX1
GO:0005739 mitochondrion
IDA
Evidence code
Paragraph 4
The more Hex
transfected into
the cell, the lower
the luciferase
activity
Worked example 4
Hex
Cells synthesis GAL4-Hex
fusion protein
Hex
Five copies of the GAL4-binding
site upstream of promoter
Luciferase transcription inhibited,
no luciferase activity
These results show that GAL4-Hex fusion protein represses luciferase
expression by binding specifically to the GAL4-binding sites, suggesting
that Hex functions as a transcriptional repressor.
Worked example 4
Worked example 4
Look at the child terms and see
if you can find a more specific
GO term
Look at the synonyms. GO uses
‘negative regulation’ rather than
‘inhibition’ or ‘down regulation’
Worked example 4
Select this term and look
again at definition and
child terms
Plasmid DNA was used so
this looks OK
Worked example 4
Select this term and look
again at definition and
child terms
Do you know what RNA
polymerase II does? The
Henry Stewart talks will
inform you
Worked example 4
Worked example 4
GO term
Supporting
evidence
GO:0000122 negative
regulation of transcription from
RNA polymerase II promoter
IDA
Table 1, luciferase assay
shows Hex suppresses
GAL4 mediated transcription
Evidence
code
(Biological Process)
GO:0003714 transcription
corepressor activity
(Molecular Function)
IDA
Table 1, luciferase assay
shows Hex suppresses
GAL4 mediated transcription
What can we get from experiments using ‘mutant’
Example
5
proteins?
GO does not annotate
protein
variants/mutation
details
Wild type SKAP55 transfected into the cell, increases
the luciferase activity
.
Worked example 5
Cells synthesis Myc-SKAP55 fusion
protein
IL-2-Luciferase transcription enhanced by
SKAP55
This experiment is different from the previous one
because the SKAP55 is not linked to a domain that binds
DNA
Could SKAP55 affect a different part of the signaling
pathway?
T cell Receptor Signalling
IL2
IL2R
IL2
Anti-CD3+PMA+iono
SKAP55?
Worked example 5
Mimic TCR activation
SKAP55 expression increases
transcription from IL2 promoter
No Molecular Function annotation
GO term (Biological
Process)
GO:0045944 positive
regulation of transcription from
RNA polymerase II promoter
GO:0050852 T cell receptor
signaling pathway
Supporting evidence
Figure 7A luciferase assay shows Wild type
SKAP55 increases transcription from IL-2
promoter
Figure 7A. anti-CD3 antibody plus PMA and
ionomycin mimic stimulation of the T cell receptor
(TCR). Over expression of SKAP55 increases the
response to TCR stimulation
Ev.
code
IDA
IMP
IDA
IMP
Annotation Example 7
Which protein to annotate?
Figure 1 from Kim M.L. et al PLoS One. 2010 Oct 15;5(10):e15371
Figure A
NF-kB translocates to nucleus
in response to S. flexneri
infection.
What annotations can we associate with NF-kB?
NF-kB: nucleus, cytoplasm
Figures B,C & D
It is the ‘target’ of NOD1 activity, NF-kappaB is imported to the nucleus, it’s
Use of NOD1 siRNA to
activity does not regulate this process
What annotations can we associate with NOD1?
NOD1 protein, inhibits NF-kB
translocation to nucleus in
response to S. flexneri
infection.
Thus NOD1
a role in
translocation
of NF-kB in response to
NOD1:
positivehas
regulation
of nuclear
NF-kappaB
import into nucleus
S. flexneri infection.
Annotation Example 7
Supporting evidence
Ev.
code
Protein
GO term
p65
NFKB
GO:0005737 cytoplasm Figure 1A immunoflorescence
shows cytoplasmic location of
NFkB in uninfected cells
p65
NFKB
GO:0005634 nucleus
Figure 1A immunoflorescence IDA
shows nuclear location of NFkB
in infected cells
Nod1
GO:0042346 positive
regulation of NFkappaB import into
nucleus
Figures 1B,C,D show depletion IMP
of NOD1 by RNAi, inhibits
NFkB translocation into the
nucleus in infected cells
IDA
Note that different evidence codes are used for a single paper